Cost of Care for Newborns With Neonatal Abstinence Syndrome in a State Medicaid Program

This cohort study estimates 10-year costs of care for neonatal abstinence syndrome in a neonates with Medicaid-eligible mothers in Alabama.


+ Supplemental content
Author affiliations and article information are listed at the end of this article.involving the central nervous and gastrointestinal systems in newborns.A prior study showed that 98% of neonates with NAS were exposed to opioids in utero. 1 Neonates with NAS are more likely to have birth complications, including low birth weight, breathing issues, jaundice, feeding difficulties, seizures, and possible sepsis. 2,3While not often fatal, NAS generally involves extended hospital stays after birth.Onset is typically within 60 hours after delivery and can last from 2 days to over a month. 4e incidence of NAS has been increasing throughout the US. 5 Incidence in the East South

Open
Central US census region (Kentucky, Tennessee, Mississippi, and Alabama) is the highest in the US, with 16.2 births of neonates with NAS per 1000 population as of 2015. 3 In Medicaid populations, the incidence and cost of births of neonates with NAS are also increasing. 4 Using 2014 data, Winkelman et al 6 estimated that NAS-related birth expenditures cost Medicaid about $462 million annually.Corr and Hollenbeak 7 estimated that these expenditures were $316 million in 2012 and attributed at least some of the increase to increasing incidence rates of NAS.Alabama is 1 of 16 states not currently participating in the Healthcare Cost and Utilization Project, which many recent studies have used to examine NAS costs and incidence rates. 8deral and US state governments handle most of the economic burden associated with NAS, as Medicaid programs financially cover most neonates born with NAS in the US. 3,7In Alabama, Medicaid covers over half of deliveries. 9The population of neonates in Alabama whose births were paid for by Medicaid has experienced trends analogous to those in the literature, including an increasing number of neonates born with NAS. 4,5Additionally, Alabama had the highest rate of opioid prescriptions per capita between 2019 and 2021. 10e economic burden of NAS does not immediately dissipate after birth, as associated issues likely have long-term consequences. 10,11Neonates diagnosed with NAS in commercial insurance populations are more likely to use health care resources at a higher rate later in life compared with those without NAS. 9This increased use may be associated with a long-term cost burden.
Furthermore, children with prenatal opioid exposure have more regulatory problems and delays in cognitive functioning than those without. 11,12st-related NAS inquiries among Medicaid populations have primarily relied on the publicly available Kids' Inpatient Database (KID), which is not structured for studies with longitudinal designs. 3,7Moreover, KID data are only available every 3 years and, therefore, are unsuitable for examining incidence rates at policy-relevant time intervals.A long-term study by Corr et al 13 analyzed 11 years of data using the Medicaid Analytic eXtract files but had issues similar to those of studies using the KID data, 3,7 as the most recent year of these data are 2013.On the other end of the spectrum, Liu et al 14 were able to examine long-term health care use and cost but used data provided by commercial insurance providers to calculate cost ratios. 10Their insights offer understanding of the total economic burden of NAS.Still, given that most of this burden lies with state Medicaid populations, many unknowns remain regarding NAS-associated health care costs in the US.
Moreover, Medicaid and commercial insurance populations vary significantly in sociodemographic characteristics and opioid prescribing practices for insured recipients. 14,15 aimed to estimate the cost of care for neonates born to Medicaid-eligible mothers who were diagnosed with NAS within 30 days after birth.To this end, we considered the costs for individuals born with NAS in Alabama's Medicaid population from birth to the end of continuous eligibility or the 10th year of life, whichever came first.Additionally, we aimed to estimate the distribution of these financial consequences for Medicaid.
Given the insights of existing literature, [9][10][11] we hypothesized that health care costs for individuals born with NAS would continue to increase as they aged.While cost differences between individuals born with vs without NAS are smaller in the period following birth than long term, we expected them to differ statistically.Liu et al 14

Design
We designed this cohort study to investigate the long-term economic burden of a diagnosis of NAS after birth.Methods were chosen to estimate the relative costs of each included variable and to calculate NAS-related costs for individuals born in the study period.Our design was a derivative of the methods used by Corr and Hollenbeak 7 to estimate health care utilization among children with NAS.Given our similar goal of estimating cost and our equivalent risk of confounding, we also chose to run regression on a matched sample. 15

Data Sources
Data for this study came from the Alabama Medicaid Agency's administrative database.The database includes reimbursed health care claims and sociodemographic information at the individual level between January 1, 2010, and December 31, 2020.
In selecting members to be included in the study, we chose to only include individuals with at least 1 year of eligibility following birth and whose mothers were also enrolled in Medicaid at the time

Independent Variables
Sociodemographic and health-related variables were considered in this study.Following the design of Corr et al, 13 we included sociodemographic and potential confounding factors in our analysis.

Statistical Analysis
Our statistical analysis plan closely followed that laid out by Corr and Hollenbeak to the confinement of the study population to Alabama and the similar socioeconomic status of participants given Medicaid enrollment, these elements were not included in matching.
Next, we performed multiple regression on the unmatched and matched samples to separate the respective costs of NAS over the term of eligibility.We used a generalized linear model assuming a γ distribution and a log link function.This model has become a standard tool in estimating health care costs due to the inherent skewness of these costs. 17,18Marginal effect estimates are reported to show the cost of each variable.To assess the cost distribution over the 10-year period, we aggregated paid amounts over the years following birth.We then performed t tests between the groups with and without NAS for years 1 through 10.A Bonferroni adjustment was used to avoid inflating type I error.
Analyses were conducted in June 2022 using Stata, release 16 (StataCorp LLC). 19Matching was performed using psmatch2 and regression using glm in Stata, release 16. 20

Results
Among  We were also interested in how this cost was distributed over time.Given that $18 000 of total costs can be associated with NAS and $12 000 can be attributed to the initial hospital stay, 7 we inferred that NAS would account for long-term spending.Significance tests showed that individuals

Discussion
To our knowledge, this is the first study to use data more recent than 2014 in examining long-term costs of NAS and the first to link records of mothers and neonates.Given that Alabama has the highest opioid prescribing rates in the country 10 and a rising incidence of NAS, 4 understanding of the long-term costs of NAS in a worst-case scenario allows for better planning and management of resources.Rising NAS incidence also underscores the importance of current data.Between 2010 and 2020, the percentage of neonates born to mothers covered by Alabama Medicaid increased from 0.6% to 2.0%. 9These individuals also represent a disproportionate percentage of costs, with neonates with NAS accounting for 3.3% of first-year infant payments in 2020. 9Our ability to link data from mothers and neonates adds a layer of credibility through more actionable insights.
Furthermore, there is a high density of NAS cases in the southeastern US, where medical costs and the cost of living are similar to those in Alabama. 2 Examining the Medicaid population is crucial given that state Medicaid agencies 3,7 cover approximately 80% of individuals born with NAS.This demographic profile of individuals in this study is comparable to that of the population in the existing literature.As in studies by Winkelman et al 6 and Corr and Hollenbeak, 7 the population with NAS in the current study was mainly White and males were overrepresented.The incidence of NAS in the US population is also rising. 5This is not surprising given the rise in maternal opioid use. 2 In Alabama, women receiving Medicaid benefits because of pregnancy are automatically enrolled in the Alabama Coordinated Health Network (ACHN) program. 23In response to the increase in substance use disorder and infant mortality, ACHN entities screen pregnant women for opioid use as part of the case management system.The fiscal implications of the current study suggest that neonates with an NAS diagnosis may also need to be enrolled in case management to manage longterm costs and address developmental delays before they reach the age to begin school.
Not all health care costs should be viewed as negative; use of services can lead to early-stage diagnoses or prevention of health conditions.As such, Medicaid members with no health care costs are a major concern.Most state Medicaid agencies report health statistics such as well-child visits to the Centers for Medicare & Medicaid Services.This indicates the importance of these visits to prevent additional health care utilization and costs at a later date. 24e current study's findings suggest that NAS has health consequences extending to adolescence and long-term fiscal outcomes for public insurance agencies beyond the first year of life.
This study is an important step in elucidating the consequences of NAS for the US health care system.

Limitations
This study has limitations.Identifying individuals born with NAS relies on a timely, adequately documented claim.Individuals could have NAS without an identified ICD-9 or ICD-10 code on record, or they could be diagnosed more than 30 days after birth.Propensity score matching mitigated the effects of included confounders.However, as seen in the difference in NICU days between the prematched and matched samples, the selected neonates without NAS in the matched sample needed more initial care than did those in the prematched sample.Additionally, the matching criteria may not account for all variation.As mentioned, 2014 birth weight data were missing, and these neonates were excluded from the matched sample.The costs incurred by individuals born with NAS may also vary across states and insurance programs due to unique policies and programs.
Furthermore, the environment for individuals with vs without a previous NAS diagnosis can vary widely, as socioeconomic status and unemployment are associated with opioid use. 268][29] These factors could also account for increased medical costs.

Conclusions
In this cohort study of neonates born into the Alabama Medicaid population, those with an NAS diagnosis had a different demographic profile than those without NAS and incurred higher costs to state Medicaid agencies.These findings suggest that further efforts are warranted to reduce the occurrence of NAS.

Findings
Neonatal abstinence syndrome (NAS) is a withdrawal diagnosis experienced by neonates following substance use by the mother during pregnancy.The term neonatal opioid withdrawal has also been used but falls under the broader category of NAS and refers to withdrawal symptoms directly correlated with opioid use.Herein, NAS refers to a collection of drug-related withdrawal symptoms Key Points Question What is the cost of care for neonates born to Medicaid-eligible mothers who are diagnosed with neonatal abstinence syndrome (NAS) within 30 days of birth?In this cohort study of 346 259 neonates born eligible for Alabama Medicaid, those with an NAS diagnosis within 30 days of birth had estimated long-term health care costs $17 921 higher than those without an NAS diagnosis.Meaning The findings suggest that the cost of care for individuals born with NAS places a significant burden on state Medicaid agencies and calls for the mitigation of opioid use in expectant mothers.

4 Outcomes
of birth.Linking individuals to their respective mothers provided additional birth and prenatal care information.Additionally, limiting this study to individuals with an associated mother provided a population over which Medicaid programs hold influence.We obtained NAS status using International Classification of Diseases, Ninth Revision (ICD-9) andInternational Statistical Classification of Diseases and Related Health Problems, Tenth Revision(ICD-10) diagnosis codes.These codes included ICD-9 7795 (drug withdrawal syndrome in newborns), ICD-10 P961 (neonatal withdrawal symptoms from maternal use of drugs of addiction), and ICD-10 P962 (withdrawal symptoms from therapeutic use of drugs in newborns).Individuals were determined to have had an NAS diagnosis if a claim with 1 of these diagnosis codes occurred within 30 days of birth.This cutoff was used because NAS generally persists for less than 1 month.The dependent variables of interest in this study were the total incurred cost throughout Medicaid enrollment over the first 10 years of life and the yearly distribution of this expenditure.This included costs related to the initial diagnosis and treatment of NAS.Costs were derived from amounts paid by Alabama Medicaid for claims initiated within 10 years of birth.Payments were adjusted for inflation to 2021 US dollars using the US Bureau of Labor Statistics Consumer Price Index.16

Future studies should examine
the cost and long term impact of NAS in other US Medicaid populations.Under this study's model, assuming the reference characteristics of the study population, NAS will account for $3.8 million over 10 years in spending incurred by the 212 individuals born with NAS to Medicaid-eligible mothers in Alabama in 2021.These costs could be avoided with intervention in the lives of expectant, Medicaid-eligible mothers.Given recent opioid-related lawsuits and settlements, $50 billion will be made available to various states and localities.25As administrative structures are implemented to distribute this money, interventions should be considered to combat the occurrence of NAS and reduce associated costs.25 7ociodemographic factors included sex, race and ethnicity, birth year, and whether the individual lived in a rural or urban environment at birth.Sex was treated as a binary variable.Race and ethnicity, included for consistency with the study by Corr and Hollenbeak,7were self-reported and categorized birth and could indicate additional health problems later in life.Both were treated as continuous variables.As birth weight data were not available for neonates born in 2014, these neonates were removed in the matching process.The presence of a previous opioid prescription was treated as a binary variable.The number of Medicaid eligibility months was treated as a continuous variable to account for enrollment variation.Birth weight was the only variable with missing data.
as Black, Hispanic, White, and other (which included all non-Hispanic races not listed).The birth year and the mother's age at delivery were treated as continuous variables in the main model.Gestational age at birth, birth weight, and the presence of a previous opioid prescription for the mother were considered factors associated with health.Gestational age (weeks) and birth weight (grams) areJAMA Network Open | Substance Use and AddictionCost of Care for Newborns With Neonatal Abstinence Syndrome in a Medicaid Program JAMA Network Open.2024;7(2):e240295.doi:10.1001/jamanetworkopen.2024.0295(Reprinted) February 26, 2024 3/10 Downloaded from jamanetwork.comby guest on 03/11/2024 indicators of premature 7in their 2017 article given the similar goal in estimating the economic impact of NAS.As such, we conducted matching and multiple regression to estimate long-term costs.We performed t tests on continuous variables and χ 2 tests on categorical variables to assess the variation related to an NAS diagnosis.Propensity score methods were used to match individuals with and without an NAS diagnosis based on sex, rurality, birth year, eligibility months, race and ethnicity, birth weight, gestational age, and the mother's age at delivery.These variables were selected based on the study by Corr and Hollenbeak7to account for cost variations based on demographic disparities and medical conditions likely to derive from other sources, particularly premature birth.Eligibility months helped control for variations in health care costs based on the enrollment period.As in the study by Corr and Hollenbeak, 7 nearest-neighbor matching was used at a 1:1 ratio without replacement or caliper.Due 346 259 neonates with up to 10 years of Medicaid eligibility born in the state of Alabama from with NAS, pregnancies for neonates with NAS were more likely to be high risk.Furthermore, the mean (SD) number of days spent in a neonatal intensive care unit (NICU) was higher for neonates with NAS compared with those without NAS (16.2[22.3]vs2.5[13.3]days).The characteristics of the matched population were largely similar, with slight variations in race and ethnicity and birth year distributions.Table2also shows a large discrepancy in the mean (SD) number of NICU days at 15.9 (22.2) for neonates with an NAS diagnosis and 4.9 (18.0) for population without NAS (48.2%, female; 50.5%, male; and 1.3%, unknown).These disparities are detailed in Table1.Neonates with an NAS diagnosis had a lower weight at birth (mean difference, −212.0 g; 95% CI, −231.1 to −192.8 g), and mothers of neonates with NAS were a mean of 3.4 years (95% CI, 2.6-4.2years)older at delivery.While pregnancy risk status was not included in the matching criteria becauseJAMA Network Open | Substance Use and AddictionCost of Care for Newborns With Neonatal Abstinence Syndrome in a Medicaid Program JAMA Network Open.2024;7(2):e240295.doi:10.1001/jamanetworkopen.2024.0295(Reprinted) February 26, 2024 4/10 Downloaded from jamanetwork.comby guest on 03/11/2024 of direct correlations

Table 3
Each additional month of Medicaid enrollment was estimated to cost $126.The same individual would have an additional cost of $4504 if they were born 3 weeks earlier than the mean gestational age at birth, calculated by multiplying the cost associated with gestational age at birth (−$1501) by 3

Table 1 .
Characteristics of Neonates With and Without NAS Cost of Care for Newborns With Neonatal Abstinence Syndrome in a Medicaid Program and subtracting it from the total cost.Individuals with an NAS diagnosis had an estimated health care cost 2.3 times as much as those without.An NAS diagnosis accounted for an additional $17 921 (95% CI, $14 830-$21 012) in estimated paid amounts over 10 years.

Table 2 .
Characteristics of Neonates With and Without NAS After 1:1 Matching c Includes all non-Hispanic races not listed.

Table 3 .
Marginal Effect Estimates From a Generalized Linear Model of 10-Year Cost For the reference population, it was assumed that the neonate would not have NAS.If the infant did have NAS, the cost would increase by an estimated $17 921.
a In 2021 US dollars.b

JAMA Network Open | Substance Use and Addiction Cost
22 Care for Newborns With Neonatal Abstinence Syndrome in a Medicaid Program These higher costs likely correlate with children starting school, where additional health and development problems can be discovered.22Table4 depicts cost differences across the first 10 years of life.

Table 4 .
Distribution of NAS-Related Costs by Year of Life a